Medical MM Leaderboard - PMC-VQA: leaderboard
The PMC-VQA slice of the Medical Multimodal Leaderboard — questions about biomedical figures collected from PubMed Central articles; accuracy (%).
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 22 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | InternVL3-38B | 56.6 | #405 |
| 2 | Gemini 2.5 Flash | 55.4 | #245 |
| 3 | GPT-4.1 | 55.2 | #248 |
| 4 | Claude Sonnet 4 | 54.4 | #199 |
| 5 | InternVL3-14B | 54.1 | #505 |
| 6 | InternVL3-8B | 53.8 | #620 |
| 7 | Qwen 2.5 VL 7B | 51.9 | #643 |
| 8 | MedGemma-4B-IT | 49.9 | #853 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=medical-mm-leaderboard-pmc-vqa · How It Works · Data refreshed daily, snapshot 2026-10-20.